Robust face recognition after plastic surgery using region-based approaches

• We compare region-based approaches in addressing face recognition after plastic surgery.
• We exploit a split face architecture to address face recognition as a multibiometric problem.
• We propose collaboration among component modules of a multibiometric system.
• We implement a supervisor module to coordinate feedback from any response.

This paper advances the use of region-based strategies for addressing the problem of face recognition after plastic surgery. The proposed methods implement the region-based approach in several ways. FARO (FAce Recognition against Occlusions and Expression Variations) divides the face into relevant regions (left eye, right eye, nose and mouth) and then codes them independently using Partitioned Iterated Function System (PIFS) processing. FACE (Face Analysis for Commercial Entities) applies a localized version of image correlation index. Finally, the Split Face Architecture (SFA), adaptive and integrative in nature, can leverage any known recognition method, from PCA to most recent ones (including FARO and FACE), provided that it is possible to divide the face into regions. Experimental results, compared with those available from recent experiments reported in literature, show that our methods yield much better performance than state-of-the art algorithms, both holistic and region based.